README_HF.md

metadata

title: Alpha-Index 100
emoji: 📈
colorFrom: blue
colorTo: green
sdk: gradio
sdk_version: 3.50.2
app_file: app.py
pinned: false
license: mit

📈 Alpha-Index 100

ML-Powered Private Markets Investment Index

An interactive machine learning application that predicts private equity fund performance using XGBoost and advanced feature engineering.

🎯 What It Does

This app helps you evaluate private equity funds by predicting their likelihood of achieving top-quartile performance based on:

• Investment Strategy (Buyout, VC, Growth Equity, Private Credit, Real Estate)
• Fund Size ($50M - $10B)
• Vintage Year (2010-2023)
• Macroeconomic Conditions (Interest rates, Market valuations)

🚀 How to Use

1. Select your fund parameters using the sliders and dropdowns
2. Click "Score This Fund" to get ML predictions
3. View the results with conviction levels:
   • 🟢 HIGH CONVICTION (≥70%): Strong top-quartile potential
   • 🟡 MODERATE (50-70%): Decent potential
   • 🔴 LOW CONVICTION (<50%): Below threshold

🤖 Technology

• ML Model: XGBoost Classifier
• Features: 11 engineered features including interaction terms
• Performance: ~85% ROC-AUC on test data
• Training Data: 5,000 synthetic fund records

📊 Model Features

The model considers:

• Fund strategy and size
• Market conditions at fund launch
• Interest rate regimes
• Valuation regimes
• Interaction effects between variables

⚠️ Important Note

This application uses synthetic data for demonstration purposes. All fund names, performance metrics, and relationships are artificially generated. This is an educational project showcasing ML capabilities in private markets.

🔗 Project Repository

Full source code, data generation scripts, and model training pipeline available in the repository.

---

Built with ❤️ using Gradio and XGBoost